working on a in house fem software. So I thought let me enumerate some
points I noted while dealing with these analysis softwares.
1. Understand and be comfortable with the fact that GUI is not
important.
Most of the GUIof analysis programs suck.
And it doesn't matter because in the long run, the solver is the one
where most of the analysis results should come from.
2. Everything can run though batch.
As I said GUI is just an addin, for efficiency and automation,
knowing how to run the softwares to run in batch is important. This is
where you should spend your maximum time while learning. Learn how to
complete an analysis with ASCII text files.
3. Understanding the underlying principles is more important than the
technique.
How to create model, appling boundary are crutial but knowing why
those boundary conditions are needed is more important. Knowing which
element to use is more important that how to add elements.
4. Use small models.
Big complex model look sexy. Working with them provide a sense of
accomtplishment, but if you want to truly learn the priciples of
underlying theory and application then use small models. They are easy
to modify, easy to solve and most importantly easy to try out
different combinations. Basics come not from analying a big 3 million
nodes landing gear model but playing and tinkering with small cube.
5. Experiment with the core functionality of the software not the
models. As I said earlier, models are just models. You are learning
analysis here, so keep learning that, models are just a way to learn
the different features.
6. Keep the end goal in mind.
Don't forget you are learning to do analysis. It's about designing
products. You are learning it to design safe, reliable designs. So
learn with that thing in mind. Also remember automation, the better
you are with it, the more efficient analyst you will become.
7. Verify all your results.
It's a software. Your job is not just to setup experiments, and
present the results. An analyst job begins after the run. It's the
interpretation of the results that matter. Verifying what you got
after the run is what you should teach yourself.
These were some of the things I remember and will suggest to anyone
learning analysis softwares like ansys, nastran , ls dyna etc.
